Ontologies for Agent-Based Modeling?

[Michael DeBellis]

I've been taking a fascinating online course from the Santa Fe institute on Agent-Based Modeling. The idea is that you can model complex systems that have emergent behavior using ABM. The first day of class we built a simple model that creates simulations of bird flocking behavior with just a few lines of code. The class uses an open source tool called Net Logo that is very cool. But this also seems to me like a domain where ontologies would be natural. I'm also planning to look at a Python library called  AgentPy because I think it will scale better than Net Logo and of course it provides all the mathematical and data analysis libraries in Python. 

Does anyone know if there are any reusable ontologies in this domain? Note this isn't the same as the kind of Agent modeling and planning that you find in tools like PDDL (Planning Domain Definition Language... a very cool language for describing Agents, States, Tasks and plans). Although I think there might be a way to integrate the two concepts so actually, I'm interested in any Agent ontology models of that type as well. 

[Chris Mungall]

Hi Michael,

ABMs can be used in varied domains like epidemiology, ecology, economics. Ideally all of the agent types, agent properties, and system properties in an ABM would be documented using the relevant domain ontologies. For example, an infectious disease ABM could reuse concepts like R0, Case Fatality Rate, etc, as well as using standard disease identifiers from the relevant biological ontologies.

I think this is the most compelling use of ontologies as it allows models to be integrated and compared. An analogous effort would be something like CMIP6 (https://pcmdi.llnl.gov/CMIP6/) which integrates multiple climate models using a standard vocabulary (CF; https://pcmdi.llnl.gov/mips/cmip/std_20c3m.html). (they don't call this an ontology but it effectively functions as one).

However, I don't think the ABM community is at a level where coupling different ABMs is as compelling as it is for climate models. (I would like to be proven wrong here however!)

But I think you're asking for a general ontology that would be applicable across multiple ABMs? Something like the FLAME model (https://docs.flamegpu.com/guide/creating-a-model/index.html) or the NetLogo dictionary (https://ccl.northwestern.edu/netlogo/docs/dictionary.html)

Another use case would be to describe all of the contextual information and assumptions about a model

The ODD protocol is used to describe ADMs (without ontologies)

https://www.jasss.org/23/2/7.html

There is the potential to use more formal representations here, but the article mentions difficulties getting even less formal conventions adopted. But of course LLMs can be used to help with this task.

Some more philosophical articles:

https://link.springer.com/chapter/10.1007/978-3-540-70556-7_11

https://www.jasss.org/13/1/3.html

CellML, SBML, BioPAX
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2728449/

On Tue, Jun 25, 2024 at 11:10 AM Michael DeBellis <mdebe...@gmail.com> wrote:

I've been taking a fascinating online course from the Santa Fe institute on Agent-Based Modeling. The idea is that you can model complex systems that have emergent behavior using ABM. The first day of class we built a simple model that creates simulations of bird flocking behavior with just a few lines of code. The class uses an open source tool called Net Logo that is very cool. But this also seems to me like a domain where ontologies would be natural. I'm also planning to look at a Python library called  AgentPy because I think it will scale better than Net Logo and of course it provides all the mathematical and data analysis libraries in Python. 

Does anyone know if there are any reusable ontologies in this domain? Note this isn't the same as the kind of Agent modeling and planning that you find in tools like PDDL (Planning Domain Definition Language... a very cool language for describing Agents, States, Tasks and plans). Although I think there might be a way to integrate the two concepts so actually, I'm interested in any Agent ontology models of that type as well. 

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/2f26c432-dbc3-481b-a28f-0fb65123a19an%40googlegroups.com.

[Michael DeBellis]

Thanks Chris, Outstanding!  that is all very useful I appreciate it. At least for now, climate models, actually any models relevant to modeling issues related to climate change like the Carbon cycle, the hydrological cycle, biodiversity models, disaster models (chemical spills, hurricanes) etc. are what I'm interested in. I've just started getting into ABM but I really like it and think there is a lot of potential to use ontologies with ABM. Will check out those links. 

Michael

You received this message because you are subscribed to a topic in the Google Groups "ontolog-forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ontolog-forum/VZLRVlfC9ms/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAN9Aiftj2rJxPR-vDOVmJJeDAyoAbmTqRgApibP9-U%3D1mDxAuw%40mail.gmail.com.

[Chris Mungall]

Great!

Of course ENVO will have a lot of the processes you describe (https://obofoundry.org/ontology/envo)

I'm also very interested in integrating other kinds of models like land system models (similar to climate models but encompasses a very broad range of parameters, beyond what is in CF). We are just getting going on a project that links ontologies and land system models here at LBNL, as a proof of concept we semi-automatically constructed an ontology using LLMs derived from FORTRAN simulations: https://github.com/bioepic-data/ecosim-ontology

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CALGFikcymzVY2hD8H-K-NeiupUPCBh9BO7jcvLziGR9AZ3LyRg%40mail.gmail.com.